Introduction: London in A.D. 43

4. London in A.D. 43.

South-eastern Britain was overrun by the troops of the Roman emperor Claudius in A.D. 43. Whether London was founded under the new régime, or whether the Roman city developed round a pre-Claudian nucleus, is an old problem which is still a new one. The discussions to which it has given rise have tended to converge upon (i) the a priori likelihood that an open, well-watered, accessible and fairly commanding site such as that of London would have attracted prehistoric settlers; and (ii) the fact that the name of London, however interpreted, (fn. 1) is of Celtic origin, with the inference that the settlement must therefore have been a "Celtic" foundation. Like all such arguments, the former can only stand if supported by tangible evidence, whilst the latter is nullified by the well-known preference of the Romans for native names, even for frontier-posts planted by them on previously uninhabited sites. Haverfield, whose own conclusion was that "either there was no pre-Roman London, or it was a small and undeveloped settlement, which may have been on the S. bank of the Thames," refers in this connection to the frequency with which modern Europeans have used a native nomenclature for towns which they have themselves founded in their colonies; he might to-day have cited Canberra as an apposite illustration. Before accepting Haverfield's qualified rejection of a pre-Roman London, however, it is desirable to review the evidence in the light of the discoveries and experience which have accumulated during the sixteen years since he wrote.

The literary evidence, quoted in Section 1 (p. 1), is clear on one point: Tacitus, writing of circumstances within the experience of his own generation, states that in the year 60 London was a busy and populous trading-centre. Cassius Dio may be thought to carry us farther back. In his description of the Roman advance to the Thames in the year 43 he refers to difficult fords near the point where the river "empties into the ocean and at flood-tide makes a lake." The reference can scarcely be other than to the fords which, at favourable states of the tides, then lower than to-day, formerly existed between Battersea and the Pool. But, after recounting the initial failure of the Romans to follow the retreating Britons across these fords, Dio suddenly introduces a bridge "a little way up stream." Whether this bridge was a light raft-construction put together by the Roman engineers, or whether it already existed but was ignored in the preliminary stages of the rout, is as obscure to us as it may have been to Dio himself writing 180 years or more after the event. Unfortunately, a point which was of little moment to Dio is of vital importance to us; for a pre-Claudian bridge in the neighbourhood of London would have implied the existence of some sort of bridge-head settlement there before the invasion, and would thus have gone far to solve our initial problem. As it is, Dio's statement cannot be claimed as evidence, and it is necessary to look elsewhere—to archæology—for the solution.

No structural evidence of a pre-Roman London has yet come to light. Attempts have been made to interpret some of the numerous pile-structures which are found from time to time in the old bed of the Walbrook as the remains of a "Celtic" settlement. Even General Pitt-Rivers, who, in 1866, closely watched the uncovering of piles in the alluvium of the Walbrook adjoining the street known as London Wall, was influenced by the discoveries of prehistoric lake-dwellings which were at the time the dominant topic in European archæology. (fn. 2) Having admitted that "it is very remarkable that .... Roman remains are interspersed at different levels from top to bottom throughout the peat" which enclosed and, at least in some cases, preceded the piles, he concludes: "That they [i.e. the habitations formerly supported by these piles] were occupied during the Roman period is evident, but it does not necessarily follow that they were of Roman origin." Two pages later this conclusion is developed:" Upon the whole, therefore, it appears not unlikely these piles may be the remains of the British capital of Cassibelaunus, situated in the marshes, and of necessity built on piles." As we read, we are present at the birth of a legend, which was later to reach maturity in the work of Sir Laurence Gomme. This writer, having incidentally transferred General Pitt-Rivers' discoveries to "the junction of the Fleet river with the Thames," states that "the earliest London is the home of the Celtic pile-dwellers," and that "the Thames was undoubtedly the site of lake-dwellings of the familiar type made known to us principally from the discoveries in the Swiss Lakes." (fn. 3) How unworthy these piles were to carry so imposing a historical superstructure has been amply demonstrated not only by Pitt-Rivers' own record but also by the consistent evidence of similar discoveries elsewhere in London; notably those observed by Mr. F. W. Reader in Finsbury, within a stone's throw of the LondonWall group. Mr. Reader has shown that there again the evidence was overwhelmingly in favour of a Roman date for the structures. (fn. 4) But the legend will die hard.

In the absence of recorded structural remains of a prehistoric London, it is necessary to consider the evidence of minor relics from the site—particularly of the potsherds which are the surest sign of settled occupation. Metal objects of pre-Roman date are very few and scattered, and do not in themselves prove other than occasional visitation in the Bronze and Early Iron Ages. (fn. 5) In range and quantity the ceramic evidence is even more restricted. Amongst the great masses of pottery saved from building-excavations, which are now normally carried well into the natural gravel, there appears to be no single sherd of Bronze-Age or Hallstatt date or of La Tène I and II. There is, however, a small quantity of pottery which shows affinity with native types in use at the close of the prehistoric period; and the difficult problem whether this pottery represents a native occupation of the site just before the conquest, or whether it was derived from native sources in the early years of the Roman occupation, can only be discussed on a detailed survey of the evidence. This survey must be extended to include an examination of such Italic pottery as may seem to antecede the Claudian invasion and so be presumed to have been introduced by trade into a pre-conquest, but already partially Romanized, London. These two groups will be considered in turn.

A. Pottery suggesting pre-Roman types. (Fig. 2).

The most important series under this heading is a number of fragmentary pedestalurns which have come to light in various parts of the City, mostly since 1924. The pedestal-urn—a type first described by Sir Arthur Evans (fn. 6) and subsequently reviewed by Mr. J. P. Bushe-Fox (fn. 7) —is distinguished, as its name implies, by a slender and graceful pedestal-base which spreads to a well-marked disc-like foot. The lower part of the pedestal is usually made separately and joined to the main body of the vessel before baking. In earlier examples the foot is commonly hollowed, but later it tends to become flat and heavy. The body is characteristically pear-shaped, and is often decorated with raised mouldings or cordons. The clay is normally fine and well-baked, and the wheel is invariably used. The form is derived from Illyro-Italic prototypes of metal, which were copied by the potters of the Aisne and Marne in La Tène I and II (principally 4th to 2nd centuries B.C.), and later modifications of these copies were introduced by immigrants into the south-eastern counties of England during La Tène III, probably not long before 50 B.C. The occasional discovery of pedestal-pottery in association with Roman wares shows that the general type sometimes survived in this country into the second half of the first century A.D.; and it is necessary therefore to hesitate before using any individual example as evidence of pre-Roman occupation. As none of the London fragments has hitherto been published, it is desirable to examine them here in detail. They are all preserved in the Guildhall Museum.

(1) From the site of the old General Post Office in St. Martin's-le-Grand. Found in the centre of the eastern part of the site in 1925, apparently in one of a group of Roman rubbish-pits. Fragmentary pedestal of somewhat coarse grey ware. The rather high, hollow foot is characteristic of the earlier rather than the later examples of the type, but this feature alone is not a certain criterion of date since it occurs at Colchester almost certainly after the Roman conquest. (fn. 8)

(2) From the same site, S.E. corner, 1925. Fragmentary pedestal of grey ware. The very slightly hollowed base is a late feature. Beyond ascribing the specimen to the last phase of the pedestal-series, however, nothing can be said as to the date. The pedestal is said to have been found with amphorae having straight handles of an early but not closely dated type.

(3) Found with No. 2. Fragmentary pedestal of grey ware. The form of the base is very similar to that of an urn found at Folkestone with a Samian cup (form 27) and a Roman jug of fairly late lst-century type (Swarling, Pl. V, 1, and Pl. XI, 6). It is also comparable to a base found at Margidunum with Roman pottery (F. Oswald, Journ. Rom. Studies, XIII, 115).

(4) From the site of No. 112 Fenchurch Street, 1925. Fragmentary pedestal-like base of grey ware, possibly of a butt-beaker rather than of an urn of true pedestal-type. The flat base is in itself a late feature, and the base-ring is a Roman rather than a native characteristic. But the important feature of this base is that, together with No. 5, it was found with lst-century Samian pottery, including two sherds of "Ritterling form 12" and a piece of form 29 of Nero-Flavian type. The authority for this association is Mr. Q. Waddington, of the Guildhall Museum, who was present at the discovery.

(5) Found with No. 4, together with post-Claudian pottery. Fragmentary pedestal of grey ware. The high hollow foot is, at first sight, an early feature, but the upward "kick" in the centre is a Roman rather than a "Celtic" characteristic. (Compare the pedestals found with Roman pottery at Margidunum, Journ. Rom. Studies, XIII, 115.) In the early (native) pedestal-urns, the base of the main body of the vessel tends, on the other hand, to sag into the pedestal.

(6) From the site of the Bank of England, 1926. Fragmentary pedestal of dark grey ware with smooth black surface. The fabric is identical with that of the so-called "Belgic" wares of the first half and middle of the 1st century A.D. The very sharp and elaborate grooving is also quite foreign to the pedestal-pottery of the AylesfordSwarling series, and links the example with those from the earlier Roman levels at Margidunum (cited above).

(7) From the same site, 1926. Fragmentary pedestal of dark grey ware, with flat base marked by a prominent moulding. This moulding may be compared with that on a pedestal found in Colchester and "in pure Roman technique .... hardly earlier than the Claudian invasion of A.D. 43" (Swarling, 21 and Pl. X, 5). The London pedestal is even more closely similar to that of a large jug found at Silchester, and probably of mid or late lst-century date (T. May, Silchester Pottery, Pl. LXXI, 163).

(8) From the same site, 1926. Fragmentary base of dark grey ware probably, like No. 4, part of a butt-beaker. A late type; the base-ring and general fabric almost certainly indicate a Romano-British date.

(9) From Miles Lane, near London Bridge, 1926. Fragmentary pedestal of black ware with a lustrous graphite surface. For the high pedestal with central "kick," see No. 5 above. Probably after rather than before the Claudian invasion. The graphite surface is unusual; it is a late Hallstatt technique which is found in this country at Hengistbury Head, Hampshire, where one fragment was found in a deposit "probably not later than the 2nd century b.c." (J. P. Bushe-Fox, Excavations at Hengistbury Head, 44). On the other hand, this ware occurred in the hut-circles at Tilbury with abundant pottery of c. A.D. 50 and later, and has been found by the present writer in rubbish-layers on the site of the Bank of England with Roman pottery of c. A.D. 50–100.

(10) From the same site, 1926. Graphite-coated black rim with smoothed linear pattern; probably part of the same vessel as No. 9. A somewhat weak and flattened variant of the pedestal-urn rim.

(11) From the same site, 1926. Fine graphite-coated black rim with ornament of applied dots. This rim can be matched in the Swarling and related series.

(12) From the site of Fenchurch Buildings, Fenchurch Street, 1924. Fragmentary pedestal of smooth grey ware. The flat base and elaborate grooving are again foreign to the Swarling pedestal-series and suggest Romano-British workmanship.

(13) From the South-Eastern Railway, Southwark, 1864. Biconical vessel of smooth grey ware, with pedestal-foot and sharply defined girth-mouldings. This vessel, illustrated also in the Guildhall Museum Catalogue, PL VI, 3 (note also 4, 5 and 10 on that plate), is here chosen as the outstanding example of a considerable quantity of London pottery which shows marked "Celtic" characteristics but cannot at present be ascribed to a pre-Claudian date. Types akin to the present example, but less finely moulded, occur at Silchester with Roman pottery (May, Silchester Pottery, Pl. LXXIX, 14; cf. Pl. LXXI, 166), and at Claudian Hofheim. The form is developed from that of the tazza or pedestal-bowl, such as occurs with the normal pear-shaped pedestal-urn at Shoebury and Welwyn. In origin, the tazza consisted of a small, sharply carinated and cordoned cup, to which a high pedestal had been added for convenience. The definitely pre-Roman examples generally emphasize the essential independence of cup and pedestal, but derivatives from Roman sites tend to amalgamate them by making the cup biconical or curvilinear and so easing the transition of its lines into those of the pedestal. The Southwark example well illustrates this process, in an advanced stage.

To review the evidence of these thirteen examples—the only examples of London "coarse" pottery which may profitably be brought into a discussion of the present problem—it is at once apparent that, on the one hand, in no case is a pre-Claudian origin certain, whilst, on the other hand, there are in most cases features which definitely distinguish our group from the known prehistoric series. The distinguishing features of the London examples are: (i) the flat base, or, where the pedestal is hollow, the tendency for the top of the hollow pedestal to "kick" instead of to sag as in the Aylesford-Swarling series; (ii) the pedestal of one piece with the body of the vessel, and not made separately and joined as in the Aylesford type; (iii) the exaggerated use of sharp grooves and mouldings; (iv) the hard, well-baked fabric of most examples, with a tendency to approximate to "Belgic" finish; and (v) a general technical precision which is foreign to the known native examples and seems to imply the introduction of improved mechanism. The Swarling pottery suggests the decadence of a craft-tradition; the London pedestals in several cases suggest a technical revival, though within the range of the same general cultural tradition. This revival may have been due to commercial intercourse with the Continent in A.D. 30 or 40 under the strong régime of the "King of the Britons," (fn. 9) or in A.D. 45 under the direct rule of Rome. The latter alternative is rendered the more likely by the results of the recent excavations at Richborough which have conclusively proved that native pottery of pre-Roman technique was still being made and used on a purely Roman site after the death of Claudius; and when the so-called "Upchurch" wares of the Thames and Medway estuaries have been studied adequately from a cultural and chronological point of view, it may be found that, in the outburst of commercial activity which certainly followed the Claudian conquest of south-eastern Britain, our native crafts were re-fertilized by the introduction of new methods and possibly new craftsmen from Romanized Gaul. Certainly the London pottery which has been examined above suggests an old native tradition in a new environment; and to adopt the obvious course of attributing that new environment to the dominating historical event of the period is in all probability correct.

B. Imported pottery of Italic origin.

Most of the characteristic red-glazed pottery which is found in great quantities on Romano-British sites was made between the second quarter of the 1st and the middle of the 3rd centuries A.D. at various factories in Gaul or Germany. It is known with equal inexactitude as "Samian" or "terra sigillata," and these nick-names are used alternatively in the present report. The convenience of the latter is that it also includes a red-glazed fabric which is not of provincial but of Italic origin and was, indeed, the prototype of the provincial ware. This Italic factory is commonly known also as "Arretine," from the name of the place (Arretium, now Arezzo) where much, though not all of it, was made. The flourishing period of the industry was from about B.C. 30 to A.D. 20, and, though the potteries survived until the Flavian period, their export-trade had ceased by the middle of the century owing to the successful competition of the derivative factories in southern Gaul. The interest of the Italic or Arretine fabric in the present context is therefore obvious; for, if much of it is found at such a site as London, the probability, on general grounds, will be that it arrived before rather than after Claudius.

The evidence is stated in detail by Mr. Davies Pryce in a later section (p. 179), where it will be found that fifteen or sixteen pieces of Arretine ware from the City and Southwark are now identifiable. Of these, five are signed by known potters whose factories were already at work in the time of Augustus. But some at least of these factories were long-lived; the wares of three of them are found at Sels in the period Augustus-Caligula (up to A.D. 41), and one is represented at Grimmlinghausen after A.D. 40. Moreover, of the unsigned pieces from London, Mr. Pryce writes: "They cannot be assigned to the flourishing period of the Italic industry. Typologically, they belong to the first four decades of the 1st century of our era; one may perhaps be even a little later." Before drawing any conclusion from this evidence, therefore, it will be well to submit it, with Mr. Pryce's help, to a brief comparative survey.

The two German sites already mentioned carry the provincial distribution of Arretine up to, or even into, the reign of Claudius. To these may be added Hofheim, which was occupied from A.D. 40 to 51 and has produced two or three fragments, and Aislingen, which was occupied probably from the time of Tiberius to that of Domitian and has yielded three more. British sites are less determinate. The Colchester district, in spite of its prime importance in the generation before the Roman conquest, seems to have contributed only one fragment of Arretine ware. An outlying Roman "villa" at Pleshey, in the same county—a site which is scarcely likely to have attracted Italic pottery before the conquest—has produced a stamp of one of the Arretine potters represented in London. The evidence from Silchester is more abundant. Roman Silchester was unquestionably in some sense of native descent. Its name (Calleva Atrebatum) is conjoined with that of its tribe in such a way as to indicate that it was a native cantonal city rather than a purely Roman foundation; native coins inscribed CALLE show that it possessed some sort of a mint in pre-Roman times; and the outer system of defences, though unexplored, is probably pre-Roman and suggests that here the Roman town was actually superimposed upon the native site, which was thus occupied continuously from the later prehistoric into the Roman period. And it has produced no less than thirty-two pieces of Arretine ware. The total is twice that of London. Before instituting a comparison between the two sites, however, it is well to consider the nature of the evidence. At Silchester, the Roman city was uncovered methodically and almost completely, but no sustained effort was made to clear consistently down to the lowest (and earliest) level of occupation where, if anywhere, pottery of pre-Claudian date would be expected. In London, on the other hand, the excavations from which pottery is obtained are not methodically observed, but they have in modern times been almost invariably carried into the undisturbed gravel; and, though the recovery of potsherds is largely fortuitous, the bright red-glazed ware stands an exceptional chance of being salved. In total quantity, there is no doubt that the mass of early Roman pottery found in London and now preserved in the British, London, Guildhall and other Museums and collections, is at least as great as that recovered by the excavators of Silchester. On these general grounds, therefore, it is tempting to compare the two sites. But further investigation reveals several indeterminate factors which may be cited in support of diametrically opposed conclusions. On the one hand, Silchester, as the capital of a Belgic tribe, is likely to have been in touch with continental trade some years before the Samian factories of Gaul had begun to capture the market from the Italic potters. If, therefore, London was not founded until A.D. 43, we might expect to find a greater disparity than actually exists between the amount of Arretine ware found on the two sites. On the other hand, Silchester was hidden away in the wolds of northern Hampshire, whereas London stood at the main gateway of Britain, and admittedly owed its early prosperity to overseas commerce. Moreover, much of the London Arretine is late in type; and these facts combined may suggest that the proportion as between a pre-Claudian Silchester in the depths of the hinterland and a Claudian London at the head of the Thames estuary is not unreasonable. But until Verulam and the habitation-area of pre-Roman Colchester (wherever it may be) have been explored, it is well to base little upon a quantitative analysis. (fn. 10)

It may, finally, be urged that, on the evidence already cited from Continental sites occupied after A.D. 40, it would be reasonable to expect at least a small proportion of Arretine ware amongst the very large mass of early post-conquest pottery recovered from London. The sudden influx of a large foreign population of soldiers and others accustomed to an Italianate environment might even be expected to induce an exceptional importation of Mediterranean wares through the new or newly-developed port, and the fifteen pieces of Arretine from London may therefore require no further explanation. Mr. Davies Pryce rightly points out (p. 181), that the evidence from this source alone is not inconsistent with a limited pre-Claudian traffic; but, if it is considered in conjunction with that derived from the "coarse" wares, the probability of a pre-Claudian settlement becomes very remote.

In summary, the geographical and archæological evidence which can be brought to bear upon the possibility, or otherwise, of a pre-Roman London is as follows:—

(i) Occupation on a site such as this, hemmed in on three sides by forest and marsh at the head of a somewhat turbulent estuary, is unlikely to have developed far except under an influential and wealthy administration such as would attract regular overseas trade and could maintain costly communications with the hinterland. These conditions were fulfilled from the outset by the Roman régime, but in preClaudian times Cunobelin at Colchester could have found little use for a Thames-side port, and, if Verulam had been the deciding factor, the landing-stage and crossing would have been expected rather at Westminster than at London. Whatever influence Verulam may have had upon the early road-system, it may safely be assumed to have had little to do with the establishment of the Southwark crossing. (fn. 11)

(ii) A small series of the earlier "coarse" pottery from London is sufficiently reminiscent of distinctively pre-Roman type to demand special consideration. It has been found, however, that two of the fourteen or fifteen pieces in question were actually associated with Roman wares, whilst the remainder tend to suggest a Romanizing tendency rather than a purely native origin. Until our knowledge of these wares becomes more exact, it is undesirable to exclude the possibility that some of the pieces may have been made rather before than after the Claudian invasion, although in every case a post-conquest date may be suspected.

(iii) The Arretine or Italic ware from London is mostly late in type, and in quantity is at least markedly less than that recovered from the almost certainly pre-Roman but comparatively remote site of Silchester. Further, if A.D. 43 still be considered late for the importation of the London Arretine, it must be remembered that soon after the conquest there was a great influx into London of traders, many of whom may be supposed to have come from districts where Arretine was still in common use, and vessels made in the previous decade or earlier would in all probability, have been introduced into this country by them. Although, therefore, the Arretine pottery is at present the most admissible evidence in favour of some sort of occupation of the site of London prior to Claudius, it is far from conclusive.

The evidence as a whole, therefore, has failed to prove the existence of a native London. But, if it has left a margin of doubt, it has at least set a limit to conjecture. It has shown clearly that we have in any case no reason for suspecting the existence of a settlement on the site more than a decade before the conquest. The whole of the "border-line" evidence could be assigned comfortably to the latter part of the reign of Tiberius. Earlier than that the evidence definitely does not warrant our going. If we do indeed go thus far and assume that some of the London Arretine in fact arrived before Claudius, a further inference is perhaps suggested. Pre-Claudian, even more definitely than Claudian, Arretine in London indicates trade. Our hypothesis would therefore involve the possibility that a few prospectors from the Roman world, like those who sometimes settled in Gaul and elsewhere before the Roman conquest, (fn. 12) may have built a wharf and a warehouse somewhere near the site of London Bridge a decade or so before the legions arrived, and so have given Roman London a "running start." Romanized traders and craftsmen were certainly not unwelcome in Britain in the days of Cunobelin, whose mint employed moneyers from the Mediterranean and whose people used Roman amphoræ alongside their own native wares. A native noble, whose richly-furnished tomb was recently explored at Lexden near Colchester, had provided himself not only with much bric-à-brac from Roman markets, but even with a portrait of Augustus himself cut from a brand-new silver coin and mounted in an ornamental frame. (fn. 13) That an occasional shipload found its way up the Colne estuary, however, does not help us in the matter of London. And there is this final difficulty. On the most liberal interpretation of the very restricted evidence, the hypothetical pre-Claudian London must have been exceedingly small, and therefore presumably concentrated. But a notable feature of the sherds of Arretine and of pedestal-pottery in London is their wide distribution. The four or five pieces of Arretine of which the actual find-spot is recorded range from Southwark to N. of Bishopsgate. The pedestal-pottery covers a similarly extensive area, from Fenchurch Street to the General Post Office. This distribution is difficult to reconcile with a small pre-Claudian trading-post; but it exactly corresponds with the distribution of the pottery of the time of Claudius and Nero (see p. 28), and is readily explained therefore in a post-conquest setting. On all grounds it must be admitted that, whilst the possibility of some pre-Claudian occupation of the site of London cannot yet be finally dismissed, there is at present no valid reason for supposing that London existed prior to A.D. 43.